Study On Multi-granulation Approximate Increment Method

Many data sets are dynamically changing,and the processing of dynamited data sets plays an important role in the era of artificial intelligence.Many scholars have put forward many effective and fast methods for the research of data sets under dynamic changes,incremental calculation is one of them.In addition,in the processing of data sets,it is also one of the commonly used strategies to granulate information to establish mathematical models,and use mathematical theoretical methods for analysis and research.There are two common granulation methods: models based on single-granularity and multi-granularity data.In terms of multi-granularity data model,mainly include the multi-granularity rough set model proposed by Qian Yu hua et al.On this basis,many scholars continue to explore and extend them,and put forward incomplete multi-granularity rough set,neighborhood multi-granularity rough set,variable precision multi-granularity rough set etc,which provide theoretical support for wide application of multi-granularity rough set theory in various fields.In the multi-granularity rough set model,approximate calculation plays a key role.With the dynamic changes of the data sets,some properties of the granular structure will change qualitatively,which makes approximate calculation more difficult.How to effectively deal with the problems under the premise of ensuring time efficiency is one of the current research hotspots.Therefore,based on previous studies,this paper tries to solve the approximate update problem of multi-granularity rough set by dynamic methods to reduce the time complexity.It mainly involves the following aspects:Firstly,for the classical multi-granularity rough set,this paper discusses the updating problem of approximate sets while adding the attributes and deleting the objects.In this case,the related properties and theorems are given subsequently.Based on this,an updating algorithm is designed and its effectiveness is verified by an example.In addition,the change and theorem of related properties when attributes are reduced and the domain is refined are also discussed.Secondly,for neighborhood multi-granularity rough set,this paper discusses the change of related properties when domain refinement and attributes increase or attributes decrease.Based on this,an updating algorithm is designed and its effectiveness is verified by an example,which provides an effective mathematical theory method for approximate updating of neighborhood multi-granularity rough sets.Then,for the conceptual multi-granularity rough set,this paper discusses the change of related properties when the attributes increase or decrease based on the vector matrix using the relative correct classification rate,and gives the corresponding theorems,which provides a mathematical theoretical basis for the approximate update of the conceptual multi-granularity rough sets.Finally,for variable precision multi-granularity rough set,this paper discusses the change of related properties of variable precision multi-granularity rough sets under tolerance relation by using the relative correct classification rate.Based on the vector matrix using the relative correct classification rate,when the missing value in the system are obtained the attribute value,this article discusses the change of the related properties of the variable precision multi-granularity rough sets under the tolerance relationship.Based on this,an update algorithm is designed and the effectiveness of the algorithm is verified by comparative experiments.In this paper,based on the vector matrix,the relative correct classification rate is used to discuss the change of the related properties of classical multi-granularity rough sets and neighborhood multi-granularity rough sets with double changes in the domain and attribute,the incremental method is designed accordingly.The example shows that it can effectively avoid the processing of repeated data under dynamic changes and reduce the search area of upper and lower approximation operators under dynamic changes,thus reducing time complexity and improving algorithm efficiency.